Mental disorders are among the most prevalent conditions in the United States, and their burden for the individual and society is tremendous. Nevertheless, they are still widely under-diagnosed in community- based health care systems. One way to help primary care physicians identify and monitor mental health problems is to use self-administered patient questionnaires. There are a number of well developed instruments available, but integration into clinical practice has rarely been achieved. Although psychometric characteristics of many tools are good, they still do not meet clinical needs, and a common metric to compare results from different tools is still missing. In addition, paper-pencil questionnaires have to be scored manually, which impose a key barrier for clinical practice, as provider reports for high risk patients must be timely and selective to be effective. Responding to those problems we recently built a Computer Adaptive Test based on the Item Response Theory to assess the mental health status of patients in community based health care settings (MH-CAT). Reports can be printed instantly showing severity of depressive symptoms, self reported treatment and adherence. First evidence has demonstrated that the tool is well accepted, and provides very high measurement precision with almost no floor and ceiling problems assessing the entire continuum from elevated mood to depressive symptoms using only 3-4 items. The underlying item bank allows comparing established depression tools, like the Center for Epidemiologic Studies-Depression Scale and Mental Health Inventory-5. Within a 21/2 -year project we propose to: (1) establish an adaptive algorithm for the MH-CAT to identify depressive disorders with high sensitivity and specificity, (2) demonstrate its feasibility as a routine screening instrument in clinical practice, and (3) assess its impact for case recognition and clinical decision making. To achieve these aims we will evaluate depressed patients from a Primary Care Research Network in Indianapolis. The MH-CAT will be compared to the Patient Health Questionnaire (PHQ-9), and the Beck Depression Inventory (BDI). Two large health centers of the New York City Research and Improvement Networking Group located in underserved communities in the Bronx will introduce the MH-CAT into their routine care. Within a randomized cross-over study we will evaluate the screening success and impact on clinical decision making in comparison with the PHQ-9. All previously not-recognized positive screened patients will be assessed with the Composite International Diagnostic Interview to confirm the diagnostic classification and followed-up for three month to assess which actions have been taken. Approx. 2,500 patients will be included in the study, being carried out together with scientists from the Albert Einstein College of Medicine, Regenstrief Institute, RAND Corporation, Harvard University, and QualityMetric Inc. PUBLIC HEALTH RELEVANCE: The introduction of the Mental Health Computerized Adaptive Test (MH-CAT) will help primary care physicians to identify and monitor patients with mental health problems. The project will demonstrate that an introduction of the MH-CAT is possible with minimal burden for patients and providers. It will have the potential to improve lives of thousands of patients suffering from undiagnosed mental health problems.